Data communication networks are being developed which enable the flow of information to ever greater numbers of users at ever higher transmission rates. However, data transmitted at high rates in multi-pair data communication cables have an increased susceptibility to crosstalk, which often adversely affects the processing of the transmitted data. Crosstalk occurs when signal energy inadvertently "crosses" from one signal pair to another. The point at which the signal crosses or couples from one set of wires to another may be 1) within the connector or internal circuitry of the transmitting station, referred to as "near-end" crosstalk, 2) within the connector or internal circuitry of the receiving station, referred to as "far-end crosstalk", or 3) within the interconnecting cable.
Near-end crosstalk ("NEXT") is especially troublesome in the case of telecommunication connectors of the type specified in sub-part F of FCC part 68.500, commonly referred to as modular connectors. The EIA/TIA (Electronic/Telecommunication Industry Association) of ANSI has promulgated electrical specifications for near-end crosstalk isolation in network connectors to ensure that the connectors themselves do not compromise the overall performance of the unshielded twisted pair (UTP) interconnect hardware typically used in LAN systems. The EIA/TIA Category 5 electrical specifications specify the minimum near-end crosstalk isolation for connectors used in 100 ohm unshielded twisted pair Ethernet type interconnects at speeds of up to 100 MHz.
A typical modular jack includes a housing having a cavity therein of a size for receiving a modular plug, where the cavity is provided with a plurality of cantilevered spring contacts which correspond to a like plurality of contact terminals in the mating modular plug. The modular plug receives discrete, insulated, stranded or solid conductors in conductor-receiving channels or slots formed in a dielectric housing. Flat, blade-like metallic terminals are then inserted into individual vertically oriented slots in the housing in a generally side-by-side arrangement with contact portions thereof extending into engagement with the conductors. When the plug is inserted into a modular jack, the cantilevered portions of the terminals in the jack engage portions of associated terminals in the plug.
The characteristics of Category 5 plugs must be verified to conform with FCC standard ANSI/TIA/EIA-568-A by measuring near-end crosstalk loss between the unshielded twisted pair conductor combinations when the plug is in an unmated state, i.e., when there is no current flow through the plug. This measurement is sometimes referred to as a "terminated open circuit" or TOC test.
In an eight-position modular plug, the contacts and twisted wires are numbered from 1 to 8, from left to right with the contacts facing upward. Wires 4 and 5 form signal pair number 1, i.e., they are operatively electrically coupled in an electrical circuit, wires 1 and 2 form signal pair number 2, wires 3 and 6 form signal pair number 3 and wires 7 and 8 form signal pair number 4. In this case, the TOC test is performed on the six different twisted pair conductor/wire combinations, namely the combinations of signal pair numbers 1 and 2, 1 and 3, 1 and 4,2 and 3,2 and 4, and 3 and 4.
To conduct the TOC test, the apparatus shown in FIG. 1 is used. A 100 .OMEGA. resistor 10 is connected in parallel with the 100 .OMEGA. test leads 12 (where they connect to the wideband baluns 14) and NEXT is measured by the network analyzer 16. The measured NEXT loss at 100 MHz must be in the range shown in Table 1.
TABLE 1 Wire Pair Combination Test Plug NEXT loss at 100 MHz 1 and 2 .gtoreq.55 dB 1 and 3 .gtoreq.40 dB 1 and 4 .gtoreq.55 dB 2 and 3 .gtoreq.45 dB 2 and 4 .gtoreq.55 dB 3 and 4 .gtoreq.45 dB
In addition, for wire pair combination 1 and 3, the difference between the NEXT loss measured at 100 MHz and the NEXT loss measured at 10 MHz must be 20 .+-.0.5 dB. Additional TOC requirements for wire pair combination 1 and 3 of the test plugs include: at least one of the test plugs must exhibit NEXT loss in the range of .gtoreq. 40.0 dB to &lt;40.5 dB at 100 MHz; at least one of the test plugs must exhibit NEXT loss in the range of .gtorsim. 40.5 dB to &lt;41.5 dB at 100 MHz; and at least one of the test plugs must exhibit NEXT loss in the range of .gtoreq. 41.5 dB at 100 MHz;
Conventional modular plugs include one or more load bars for receiving the conductors in separate conductor-receiving passages. The use of load bars contributes to control of the inter-conductor capacitance in the plug. FIG. 2 shows typical TOC values measured for ten eight-position modular plugs of the same design between the pair combination 2 and 4, specifically, an RJ45 plug having two load bars terminating a 24 AWG Tinned Stranded UTP cable made by Lucent Technologies. As shown in FIG. 2, for eight-position modular plugs having the same design, TOC values can vary by as much as 40 dB between plugs (compare test plugs 1 and 10). This variation is partially due to the relatively random arrangement of the unshielded twisted pairs (UTP) of conductors in the body of the plug, i.e., in the wire-receiving channels in the plug body, which causes small changes in the capacitance between the conductors.
One way to reduce inter-conductor capacitance in a plug is by offsetting adjacent conductors. Examples of this type of plug are disclosed in U.S. Pat. No. 5,628,647 (Rohrbaugh et al.) wherein the conductors are arranged in two planar arrays spaced one above the other. The offset conductors helps lower the plug's internal capacitance but does not result in stable TOC values for plugs having the same design.
In another attempt to stabilize the capacitance in an RJ45 plug in order to obtain consistent TOC values for plugs having the same design, three plugs 20 were assembled with four load bars 22 each (FIG. 3). The plugs initially were a standard RJ45 plug manufactured by Stewart Connector Systems but modified to include four load bars, and as tested, terminate a Berk-Tek Lan-Mark-350 cable (the same cable is used in all of the TOC tests described herein unless stated to the contrary). The use of four load bars fixed the inter-conductor capacitance within the length of the body of the plug. TOC measurements were then made on each pair combination to determine the degree of TOC stability. As shown in FIG. 4, the TOC values measured on the three plugs using four load bars each had less than a 4 dB variation from plug to plug.
Although the measured TOC values for a four-load bar plug as shown in FIG. 4 exhibits less variation from plug to plug than a standard Category 5, eight-position modular plug using two load bars, the wire pair combination 1 and 3 does not always yield a TOC value that complies with the requirements of TIA/EIA-568A. Indeed, the lowest TOC value obtained in the three plugs tested is 39.8 dB between the wire pair combination 1 and 3. However, the minimum requirement for pair combination 1 and 3 is 40 dB (See Table 1) and thus these modified plugs would not pass the TOC test according to ANSI standard EIA/TIA-568-A.
With respect to NEXT values (a measure of both capacitive and inductive crosstalk) between wire pairs of plugs, it has been found that variations in NEXT values between plugs of the same design are caused at least in part by the random arrangement of the UTP wires underneath the plug's strain relief element. That is, the strain relief element in typical plugs engages with a shielded cable at a location prior to unsheathing of the cable and thus prior to insertion of the wires in positioning channels in the plug (e.g., in a load bar of the plug) and therefore, the UTP wires are arranged in the cable underneath the strain relief element in an arbitrary, random manner. It has also been found that TOC values between wire pairs also vary in view of the random nature of the arrangement of the wires in the cable below the strain relief element. In this regard, FIG. 15 shows a table of the results of tests performed on ten (10) different plugs of a model of an RJ45 Category 5 plug manufactured by the assignee hereof for both NEXT values and TOC values for all of the combinations of wire pairs (e.g., wire pair 1 to wire pair 2 is represented by 45-12). The measurement of NEXT is "de-embedded" NEXT, i.e, the crosstalk of a mating plug and jack is measured and the crosstalk of the jack is subtracted therefrom so that the resultant value is only the crosstalk caused by the construction of the plug. FIG. 16 is a table of maximum, minimum and variation in de-embedded NEXT values based on the data in the table of FIG. 15. As seen in FIG. 16, the variation in de-embedded NEXT values (delta) ranges from 7.1 dB to 27.6 dB. FIG. 17 is a table of maximum, minimum and variation in TOC values based on the data in the table of FIG. 15. As seen in FIG. 17, the variation in TOC values (delta) ranges from 5.9 dB to 20.9 dB. It would be beneficial to reduce the extent of these variations in de-embedded NEXT values and TOC values since variations in NEXT and TOC values could result in adverse operational performance of the plug.